Brain Research 808 Ž1998. 279–293
Research report
Spinal inputs from lateral columns to reticulospinal neurons in lampreys Laurent Vinay
a,b
, Fulvia Bongianni c , Yoshihiro Ohta c , Sten Grillner c , Rejean Dubuc ´
a,d,)
a
d
Centre de Recherche en Sciences Neurologiques, UniÕersite´ de Montreal, Canada ´ C.P. 6128 Succ. A, H3C 3J7, Montreal, ´ Quebec, ´ b UPR Neurobiologie et MouÕements, CNRS, 31 chemin Joseph Aiguier, BP 71, F-13402 Marseille Cedex 20, France c The Nobel Institute for Neurophysiology, Department of Neuroscience, Karolinska Institutet, S171 77, Stockholm, Sweden Departement de Kinanthropologie, UniÕersite´ du Quebec Canada ´ ´ a` Montreal, ´ C.P. 8888, Succ. centreÕille, H3C 3P8, Montreal, ´ Quebec, ´ Accepted 11 August 1998
Abstract This study characterizes the inputs from the lateral columns of the spinal cord to reticulospinal neurons in the lampreys, using the in vitro isolated brainstem and spinal cord preparation. Synaptic responses to the electrical stimulation of the lateral columns were recorded in reticulospinal neurons of the posterior and middle rhombencephalic reticular nuclei. The responses consisted of a mixture of excitation and inhibition. They were markedly potentiated when using trains of two to five pulses, suggesting that the larger part of these synaptic responses was mediated via an oligosynaptic pathway. An early component, however, persisted when using twin pulses at 10–20 Hz on the ipsilateral side, suggesting the presence of an early mono- or disynaptic component. When increasing the stimulation strength, an early fast rising excitatory component appeared. It most likely resulted from an antidromic activation of vestibulospinal axons in the lateral tracts, which make en passant synaptic contacts with reticulospinal neurons. Responses were practically abolished by adding CNQX and AP5 to the Ringer’s solution. The late component of excitatory responses was decreased by AP5, suggesting that NMDA receptors were activated. The NMDA receptor-mediated component was larger when using trains of stimuli or in Mg 2q-free Ringer’s. The application of NMDA depolarized reticulospinal neurons. The glycinergic inhibitory component was markedly increased in Mg 2q-free Ringer’s. Moreover, GABA B-receptor activation with Žy.-baclofen abolished both excitatory and inhibitory responses. Taken together, the present results indicate that ascending lateral column axons generate large excitatory and inhibitory synaptic potentials in reticulospinal neurons. The possible role of these inputs in modulating the activity of reticulospinal neurons during locomotion is discussed. q 1998 Elsevier Science B.V. All rights reserved. Keywords: Spinal lemniscus; Excitatory amino acid; Glycine; GABA; Locomotion
1. Introduction In mammals, the activity of spinal projecting neurons in various supraspinal centers is strongly modulated during active and fictive locomotion w23,24,33x. In the lamprey, the activity of reticulospinal neurons is also rhythmically modulated during both active w9x and fictive w18,19x locomotion. Dubuc and Grillner w14x showed that, at least part of this modulation originates from the spinal locomotor networks. The distribution of spinal neurons, which project to the caudal reticular formation within the brainstem and which may be responsible for this modulation, has been studied anatomically w34x. Neurons were labeled on both sides of the spinal cord, although neurons with contralater) Corresponding author. Fax: q1 514-343-6611; E-mail:
[email protected]
ally projecting axons were more numerous Ž80% of total. than neurons with an ipsilateral axon Ž20% of total.. The number of labeled neurons decreased rapidly in the first spinal segments caudal to the injection site. However, still two-thirds of the spino-bulbar neurons were located more than five segments away from the injection site. Most of them Ž84%. were located in the rostral half of the spinal cord although some were located as far as 80 segments away. These axons were running in the area of the spinal lemniscus and injections of tracers directly into the posterior rhombencephalic reticular nucleus showed that many cells projected to this nucleus. The physiology of spino-bulbar neurons has not yet been studied. However, Vinay and Grillner w30x have recorded from spinal axons, which may originate from these neurons and which ran within the lateral columns of the rostral spinal cord to project to the brainstem. These
0006-8993r98r$ - see front matter q 1998 Elsevier Science B.V. All rights reserved. PII: S 0 0 0 6 - 8 9 9 3 Ž 9 8 . 0 0 8 3 5 - X
280
L. Vinay et al.r Brain Research 808 (1998) 279–293
axons displayed rhythmic discharges during fictive locomotion, which occurred in different phases of the locomotor cycle. Since the synaptic transmission in the brainstem was blocked thereby suppressing any rhythmic descending activity, these discharges were generated by the spinal locomotor networks and transmitted to the brain via an ascending axon. This suggests that spinal neurons with
axons projecting in the lateral columns could contribute to the modulation of reticulospinal neurons during locomotion. Little is known about the brainstem termination sites of ascending axons in the lateral columns. Ronan and Northcutt w25x have used anterograde degeneration techniques combined with silver staining to reveal that some axons
L. Vinay et al.r Brain Research 808 (1998) 279–293
reach the isthmus, other the mesencephalic tegmentum, the optic tectum and the diencephalon. They also showed the presence of degenerated fibers located laterally and ventrally in the rhombencephalic basal plate. This allowed for possible contacts with distal dendrites of some reticulospinal cells, which possess an extensive dendritic tree within the basal plate w7,12,20,27,28x. The present study was carried out in order to test whether ascending axons within the lateral tracts provide synaptic inputs to reticulospinal neurons. It will be shown that microstimulation of the lateral columns evokes synaptic responses in reticulospinal neurons of both the middle ŽMRRN. and the posterior ŽPRRN. rhombencephalic reticular nuclei. Pharmacological blockers acting at selective receptor types and sub-types were used to identify the neurotransmitters and receptors involved.
2. Material and methods Experiments were performed on adult specimens of Ichthyomyzon unicuspis Ž n s 26. and Lampetra fluÕiatilis Ž n s 5. as well as on larval Petromyzon marinus Ž n s 12.. All surgical and experimental procedures were conform to the Canadian and Swedish MRC guidelines and were approved by local university committees for animal care and use. The animals were anaesthetized with tricaine methane sulfonate ŽMS-222; 100 mgrl.. They were eviscerated and all muscles were stripped away leaving the brainstem and spinal cord with their underlying notochord and cranium. The ventro-medial cranium was resected to improve trans-illumination of the brainstem through the ventral side. The preparation was fixed to elastomer silicone ŽDow Corning, Sylgard 184. at the bottom of a recording chamber filled with cold Ž4–98C. Ringer’s with the following composition Žin mM.: 91.0 NaCl, 2.1 KCl, 2.6 CaCl 2 , 1.8 MgCl 2 , 20.0 NaHCO 3 , and 4.0 glucose, bubbled with 95% O 2 –5% CO 2 to pH of 7.4 Žsee Ref. w35x. Reticulospinal neurons within the PRRN Ž n s 119. and MRRN Ž n s 19. were recorded with microelectrodes filled with 3–4 M KAc ŽFig. 1A.. Only neurons exhibiting a resting membrane potential of at least y60 mV Žmean " S.D.: y73.9 " 8.4 mV. were considered for analysis.
281
Reticulospinal neurons within the PRRN were identified as projecting to the spinal cord by recording with a glass capillary tube placed on the lateral spinal fasciculus on the ipsilateral side, while an axonal orthodromic action potential was elicited by intracellular stimulation. Alternatively, they could be identified by recording the antidromic action potential elicited by stimulation of the spinal cord. The latter procedure was also used to identify reticulospinal neurons in the MRRN, which were penetrated under visual control. No attempt was made to differentiate between the different Muller cells. ¨ Glass-coated tungsten microelectrodes Ž5–15 mm exposed tip; 0.5–2.0 M V . were used to microstimulate lateral columns on either side in the rostral spinal cord Žsegments 1–15. while synaptic responses were recorded intracellularly in reticulospinal neurons ŽFig. 1A.. The stimulating electrodes were placed on the dorsal surface of the lateral columns at the level of the lateral grey matter or at mid-distance between the spinal grey and the lateral edge of the spinal cord. The same stimulation technique has been previously used to microstimulate dorsal roots and dorsal columns in lampreys w15,16x. The selective lesions of different spinal tracts which we performed showed that the technique was selective for the pathway under study with little spread of current, if any, to the dorsal columns or to pathways on the contralateral side Žsee below, Fig. 2.. Single pulses Ž1.5 or 2 ms; 0.1–30 mA. as well as trains of two to five pulses Ž100–150 Hz. were routinely used.
3. Results 3.1. Excitatory responses to lateral column stimulation Stimulation of the lateral columns on either side, at a rostral level Ž1st–8th segment. of the spinal cord, elicited synaptic responses in all of the recorded reticulospinal neurons. The responses were predominantly excitatory although inhibitory components could be revealed or enhanced with modified Ringer’s andror at depolarized membrane potential levels Žsee below.. The responses in reticulospinal neurons within the MRRN ŽFig. 1B1–B3, top trace. and the PRRN Žbottom trace. to stimulation of
Fig. 1. Synaptic responses of reticulospinal neurons to stimulation of lateral columns on both sides. A: Experimental protocol. Reticulospinal neurons were recorded within the middle ŽMRRN. and the posterior rhombencephalic reticular nucleus ŽPRRN.. Microstimulations were applied to the ipsilateral ŽiLC. and contralateral lateral columns ŽcLC. in the rostral spinal cord Ž1–15th segment.. B: Intracellular responses of two reticulospinal neurons Žone within the MRRN and the other one within the PRRN. to the stimulation of the ipsilateral lateral column at a strength Ž0.3 mA. slightly above threshold ŽT .. Traces are averages of consecutive sweeps Žnumber of sweeps indicated above each trace.. B1: Single pulse stimulations. Note the small excitatory component following the stimulus artefact. B2: Trains of three pulses at 100 Hz. B3: Trains of five pulses. Note the marked potentiation of the response. C: Intracellular responses to the stimulation of the contralateral lateral column. Note the higher threshold intensity Ž3 mA. for the contralateral stimulation compared with that on the ipsilateral side. C1: Single pulse stimulations. C2: trains of three pulses at 100 Hz. C3: trains of five pulses. Scale bars also apply to B. D: Intracellular responses to stimulation of the ipsilateral lateral column with a stronger current Ž6 times threshold intensity.. D1: Responses of reticulospinal neurons Žsame neurons as in B,C. to single pulse stimulations. D2: Responses of the MRRN neuron to twin pulses at 20 Hz. Top trace: Four single sweeps. Bottom trace: Average of four consecutive sweeps.
282
L. Vinay et al.r Brain Research 808 (1998) 279–293
Fig. 2. Effects of specific spinal cord lesions on synaptic responses of reticulospinal neurons to stimulation of the lateral columns on the ipsilateral side. A: Experimental paradigm. Responses of reticulospinal neurons within the middle rhombencephalic reticular nucleus to stimulation of the ipsilateral lateral columns ŽiLC. were recorded before and after a lesion of either the dorsal columns Žlesion B. or the contralateral lateral columns just behind the obex Žlesion C.. B: Averaged responses before and after the dorsal column ŽDC. lesion are superimposed. C: Traces are averages of 10 sweeps taken before Žcontrol. and after the specific lesion of the contralateral lateral columns ŽcLC.. C1: Single pulse stimulations. C2: Trains of two pulses.
the ipsilateral lateral columns were potentiated when using trains of stimulation of two to five pulses Žsecond spinal segment; 0.3 mA.. An early excitatory synaptic response was evoked with stronger stimulus strengths ŽFig. 1D1, 2 mA.. These short-latency fast rising excitatory components remained unchanged at a frequency of 10–20 Hz ŽFig. 1D2; four single sweeps superimposed on top and averaged trace Ž n s 4. at the bottom.. This suggests that the synaptic responses evoked from the ipsilateral side include an early excitatory component, which might be monosynaptic, followed by a depolarization most likely oligosynaptic. Responses to stimulation of the contralateral lateral columns were elicited only with higher stimulus intensity Ž5–10 times. than those elicited by stimulation of the ipsilateral side. However, single shocks applied to the contralateral lateral column rarely evoked a significant response in reticulospinal neurons of either the MRRN ŽFig. 1C1, top trace. or the PRRN ŽFig. 1C1, bottom trace.. The responses appeared and were potentiated when
using trains of stimuli ŽFig. 1C2,C3. suggesting that the responses evoked from the contralateral side are most likely conveyed through a polysynaptic pathway. In one experiment, the stimulating electrodes were kept at the same place within the lateral columns, between the second and the third spinal segments, while recording synaptic responses elicited in seven reticulospinal neurons, four in the PRRN and the remaining three cells in the MRRN. This allowed us to compare the responses and the thresholds of activation for PRRN and MRRN neurons. No difference could be observed between these two populations of neurons although the early excitatory component, evoked at higher stimulation strengths, may be recorded with a shorter latency and a faster rising slope in PRRN neurons than in MRRN cells ŽFig. 1D1, compare the averaged traces in MRRN and PRRN.. The potentiation of the responses using trains of pulses was similar for the two populations of neurons ŽFig. 1B–C, compare top and bottom traces.. Apart from the early excitatory component,
L. Vinay et al.r Brain Research 808 (1998) 279–293
the spinal inputs from lateral columns were therefore similar for the two populations of reticulospinal cells. Selective lesions of different spinal tracts were performed to further characterize the pathway conveying inputs from the lateral columns. In three experiments, the dorsal columns were lesioned on both sides ŽFig. 2A, lesion B.; the synaptic responses elicited by the ipsilateral lateral column stimulation remained unchanged ŽFig. 2B.. A lesion of the lateral columns on the contralateral side, just behind the obex, ŽFig. 2A, lesion C, n s 2. did not modify the response elicited in reticulospinal cells of the MRRN by stimulation of the ipsilateral lateral column
283
ŽFig. 2C1, one pulse and Fig. 2C2, two pulses.. Responses disappeared after lesioning the spinal cord on the same side as the stimulating electrode Žnot shown.. These results show that responses to stimulation of the ipsilateral lateral column do not depend on projections in the dorsal columns or on the contralateral side. They are carried by axons running within the ipsilateral lateral funiculus up to the brainstem level. The synaptic relays are also located on the same side. The rostro-caudal distribution of the inputs from the lateral columns was studied by recording reticulospinal neurons from the MRRN Ž n s 2., while moving the stimu-
Fig. 3. Rostro-caudal distribution of the inputs from the lateral columns to reticulospinal neurons within the middle rhombencephalic reticular nucleus. A: Schematic representation of the stimulation arrangement. The stimulating electrode was moved rostro-caudally along the spinal cord to stimulate the lateral columns at three levels Ž1st, 4th and 8th segment.. B: Responses evoked by the stimulation at various strengths Ž2, 5 and 10 times hreshold.. Traces are averages of four sweeps. B1: Stimulations at the level of the first segment. B2: Stimulations within the fourth segment. B3: Stimulations within the 8th segment. Top trace: Single pulse stimulations. Bottom trace: Trains of three pulses at five times threshold intensity. C: Mean Ž n s 10–20 trials. amplitude Žtaken from onset to peak of excitatory post-synaptic potentials. of the response to stimulation of the lateral columns with various strengths at the three rostro-caudal levels.
284
L. Vinay et al.r Brain Research 808 (1998) 279–293
Fig. 4. Chemical transmission between axons running in the lateral columns and reticulospinal neurons. A,D: Experimental paradigm. Reticulospinal neurons were recorded within the middle ŽMRRN, A–C. or the posterior ŽPRRN, D–F. rhombencephalic reticular nucleus. Stimulations were applied to the ipsilateral lateral columns ŽiLC. and to the vestibular nerve either on the ipsilateral ŽiVIII. or the contralateral ŽcVIII. side. B1,C1,E1,F1: Responses evoked in normal Ringer. B2,C2,E2,F2: Responses evoked after exposing the brainstem to a Ringer’s solution in which the Ca2q had been replaced by Mn2q in order to block all chemical synaptic transmission. Note that responses evoked by the stimulation of the lateral columns with one, three or five pulses ŽC1,F1. disappeared in Ca2q free-Mn2q Ringer ŽC2,F2. whereas an electrotonic component persisted in response to the stimulation of vestibular afferents ŽB2, arrow, E2..
lating electrode along the spinal cord to stimulate the lateral columns at various rostro-caudal levels ŽFig. 3A: 1st, 4th and 8th segment.. The threshold ŽT . was similar at
the three sites Ž1.0, 1.2 and 1.0 mA, respectively.. Responses evoked with one pulse and low intensities Ž- 5 T. from the first ŽFig. 3B1., the fourth ŽFig. 3B2. and the
Fig. 5. Dual-component ŽNMDA and non-NMDA. glutamatergic transmission between ascending axons and reticulospinal neurons. A: Effects of NMDA and non-NMDA receptors antagonists ŽAP5 and CNQX, respectively. on the responses elicited by the stimulation of the ipsilateral lateral columns in a reticulospinal neuron within the middle rhombencephalic reticular nucleus. Traces are averages of fifty sweeps. B: Increase of the NMDA receptor-mediated component with a train of pulses. Traces are averages of 10 consecutive sweeps taken before and after adding AP5 to the Ringer solution. B1: Single pulse stimulations. B2: Trains of two pulses. B3: Trains of three pulses. B4: NMDA receptor-mediated component obtained by subtracting the AP5-treated EPSP from the control response evoked with one or three pulses.
L. Vinay et al.r Brain Research 808 (1998) 279–293
285
286
L. Vinay et al.r Brain Research 808 (1998) 279–293
eighth ŽFig. 3B3. segments were similar in amplitude ŽFig. 3C. and duration. Higher intensities, which activate more axons, evoked larger responses when applied in the first four segments than more caudally. Large responses could however be evoked from caudal levels by using a train of pulses ŽFig. 3B3, bottom trace.. This suggests that, although direct spino-bulbar axons ascending from as far as the 8th segment most likely exist, as revealed by using a single pulse ŽFig. 3B3., a significant part of the lateral
column input to reticulospinal cells from the caudal spinal cord may be relayed in the upper spinal segments or the caudal part of the rhombencephalon. 3.2. What is the contribution of the Õestibulo–reticulo-spinal system to the inputs from lateral columns? The EPSPs elicited by stimulation of the ipsilateral lateral columns Žrostral segments. in reticulospinal cells
Fig. 6. Effect of removing the voltage dependent block of NMDA channels by Mg 2q ions on the excitatory synaptic responses elicited in reticulospinal neurons within the middle rhombencephalic reticular nucleus by stimulating the lateral columns. A: Synaptic responses to the stimulation of the ipsilateral lateral column ŽiLC. with a single pulse Žexcept in A3, two pulses.. Traces are averages of four sweeps except in A3 Žsingle sweep.. Responses were recorded in the control situation ŽA1., after removing Mg 2q ions from the bathing Ringer’s solution ŽA2, A3. and after washing out the effects with normal Ringer Ži.e., with a concentration of Mg 2q ions back to 1.8 mM, A4.. Note the lower sweep velocity in A3. Both the amplitude and the duration of the responses were increased in Mg 2q-free Ringer. B: Synaptic responses to the stimulation of the contralateral side ŽcLC.. B1: Control recordings in normal Ringer Ž1.8 mM of Mg 2q .. B2: Responses in Mg 2q-free Ringer. B3: Recovery of the responses in normal Ringer.
L. Vinay et al.r Brain Research 808 (1998) 279–293
may result from the activation of at least two systems: Ži. an orthodromic activation of spinal brainstem-projecting cells or axons; Žii. an antidromic activation of vestibulospinal axons which make electrotonic and chemical contacts with reticulospinal neurons w1,26x. The possible contribution of the vestibulospinal system in the effects of lateral column stimulation on reticulospinal cells was tested by replacing Ca2q in the Ringer by Mn2q Ž2.6 mM., thereby abolishing all chemical synaptic transmission. Reticulospinal responses to spinal cord stimulation were compared to those to vestibular nerve stimulation. Blocking the chemical synaptic transmission resulted in a marked decrease in the amplitude of vestibular nerve ŽcVIII.evoked responses in MRRN cells ŽFig. 4B1,B2.. A small short latency, fast rising component persisted in Ca2q free-Mn2q Ringer ŽFig. 4B2, onset of the response indicated by an arrow. and may therefore result from electrotonic transmission w26x. By contrast, the EPSP evoked by ipsilateral lateral column stimulation ŽFig. 4C1. disappeared in Ca2q free-Mn2q Ringer ŽFig. 4C2. in all the four cells tested in this way, even at a stimulus strength as high as four times the threshold Ž4T .. A larger electrotonic component was elicited by stimulation of the vestibular nerve in PRRN cells ŽFig. 4E, n s 2. than in MRRN cells. Stimulation of the ipsilateral lateral column with a low intensity Ž2T . evoked a slow rising EPSP ŽFig. 4F1, top trace. which potentiated markedly with trains of stimuli ŽFig. 4F1, middle and bottom traces.. The responses were abolished in Ca2q free-Mn2q Ringer ŽFig. 4F2.. Taken together, these results suggest that the effects elicited in MRRN and PRRN cells by stimulation of the lateral columns with a low intensity result from chemical transmission, and thus from the activation of axons issued from cell bodies in the spinal cord. Increasing the current strength reveals an early, fast rising component which may result from electrotonic transmission. The latter may take place between vestibulospinal or spino-bulbar axons located deeper in the spinal cord and reticulospinal neurons.
287
NMDA receptors in the late phase of the excitatory responses. Interestingly, the NMDA receptor-mediated component was increased both in amplitude and duration by using a train of two ŽFig. 5B2. or three pulses ŽFig. 5B3; compare the two traces in Fig. 5B4.. These data show that most of the excitation depends on non-NMDA receptor activation in normal Ringer Ži.e., with a concentration of Mg 2q ions of 1.8 mM.. Mg 2q-free Ringer was used to remove the voltage-dependent block of NMDA channels by Mg 2q ions, thereby maximizing the expression of NMDA conductances ŽFig. 6; n s 9 cells tested in this way in three experiments.. EPSPs evoked by lateral co-
3.3. Pharmacological characteristics of the excitatory responses eÕoked by lateral column stimulation in reticulospinal neurons Chemical blockers of excitatory amino acid transmission were used in order to identify the neurotransmitters involved in the lateral column pathway to reticulospinal neurons. Blocking the excitatory amino acid transmission by bath application of both NMDA and non-NMDA receptors antagonists ŽAP5, 100–150 mM and CNQX, 10–20 mM, respectively. resulted in a considerable decrease Ž80%. in the amplitude of the EPSP elicited by stimulation of the lateral columns on the ipsilateral side ŽFig. 5A.. A small residual depolarization however persisted in the presence of AP5 and CNQX Ž n s 4.. Bath application of AP5 alone slightly decreased the late part of the EPSP ŽFig. 5B1. suggesting a role of
Fig. 7. Presence of NMDA receptors on reticulospinal neurons. A: Experimental paradigm. Reticulospinal neurons were recorded within the posterior ŽPRRN. rhombencephalic reticular nucleus. NMDA Ž0.5 nl of a 10 mM solution. was pressure-ejected from the tip of a glass pipette located some 50–100 mm away from the recording electrode. Responses were recorded in the presence of tetrodotoxin Ž3 mM. added to the perfusing Mg 2q-free Ringer’s. B: Responses of the reticulospinal neuron to the pressure ejection of NMDA Žvertical arrow. before adding AP5 to the perfusing Ringer’s ŽB1. in the presence of AP5 ŽB2. and after washing out AP5 ŽB3..
288
L. Vinay et al.r Brain Research 808 (1998) 279–293
lumn stimulation on both sides were increased in amplitude and duration in Mg 2q-free Ringer ŽFig. 6A2,B2.. When two pulses were applied to the ipsilateral side, Mg 2q-free Ringer increased the initial EPSP as well as unmasked a slow component lasting 10–20 s ŽFig. 6A3, note the different time scale.. Effects were reversible and EPSPs recovered within a few minutes after adding Mg 2q to the Ringer solution ŽFig. 6A4,B3.. NMDA receptors may be located on the relay interneuronal level andror on the reticulospinal cells. Because of a previous report demonstrating the absence of NMDA receptors on reticulospinal neurons w13x, we tested this hypothesis further. The effect of local application of NMDA onto identified reticulospinal neurons within the PRRN Ž n s 5. was studied in the presence of tetrodotoxin Ž3 mM. added to the perfusing Mg 2q-free Ringer’s. As NMDA
Ž0.5 nl of a 10-mM solution. was pressure-ejected from the tip of a glass pipette located some 50–100 mm away from the recording electrode ŽFig. 7A., the membrane potential of the recorded reticulospinal cells showed a clear depolarization lasting 3 s ŽFig. 7B1; top trace.. When AP5 Ž100 mM. was added to the perfusing Ringer’s, it abolished the effect of pressure ejecting NMDA onto reticulospinal neurons ŽFig. 7B2, middle trace.. The depolarizing responses recovered after washing out AP5 ŽFig. 7B3.. 3.4. Pharmacological characteristics of the inhibitory responses eÕoked by lateral column stimulation in reticulospinal neurons Interestingly, NMDA receptors appear to also play an important role in carrying excitation to neurons which
Fig. 8. NMDA transmission occurs onto neurons which relay inhibitory inputs to reticulospinal neurons. A: Synaptic responses elicited in a reticulospinal neuron within the posterior rhombencephalic reticular nucleus by the stimulation of the contralateral lateral columns. Traces are averages of 6–12 sweeps Žnumber indicated above each trace.. A1: Responses evoked in normal Ringer Ž1.8 mM of Mg 2q . at resting membrane potential Žtop trace, note the small depolarization indicated by the vertical dotted line. or after steadily depolarizing the cell by injecting positive currents Žbottom trace.. A2: Responses evoked in Mg 2q-free Ringer. Note the large hyperpolarization after injecting depolarizing currents into the cell Žbottom trace.. B: Synaptic responses elicited in a reticulospinal neuron within the middle rhombencephalic reticular nucleus by the stimulation of the ipsilateral lateral columns. B1: Responses evoked in normal Ringer Ž1.8 mM of Mg 2q .. Four single sweeps are superimposed. B2: Responses evoked in Mg 2q-free Ringer. Note the hyperpolarization which persisted at a high frequency of stimulation Ž100 Hz, bottom trace.. C: Synaptic responses elicited in the same reticulospinal neuron by the stimulation of the contralateral lateral columns. C1: Responses evoked in normal Ringer Ž1.8 mM of Mg 2q .. C2: Responses evoked in Mg 2q-free Ringer.
L. Vinay et al.r Brain Research 808 (1998) 279–293
relay inhibitory inputs to reticulospinal neurons. In some cells, only a small depolarization, if any, could be identified following stimulation of the lateral columns in normal Ringer ŽFig. 8A1. or in Mg 2q-free Ringer’s at the resting membrane potential ŽFig. 8A2, top trace. whereas a large hyperpolarization was revealed by depolarizing Žq1 nA; about q10 mV. the cell in Mg 2q-free Ringer’s ŽFig. 8A2, bottom trace.. In other cells the hyperpolarizations could be observed even at the resting membrane potential in Mg 2q-free Ringer’s ŽFig. 8B2,C2.. As a rule, the individual IPSPs evoked from the ipsilateral side persisted at a high frequency of stimulation Ž100 Hz in Fig. 8B2, bottom trace. suggesting a very effective oligosynaptic transmission from the spinal cord to reticulospinal neurons. By contrast, the IPSPs evoked from the contralateral side ŽFig. 8C2. were likely transmitted through a polysynaptic pathway because only one compound IPSP was observed after
289
several stimuli. These results suggest that NMDA transmission occurs onto a population of inhibitory relay cells, which, in turn, inhibit reticulospinal neurons. Adding the non-NMDA receptor antagonist CNQX Ž10 mM. blocked ŽFig. 9A2. the IPSPs unveiled in Mg 2q-free Ringer ŽFig. 9A1, control; Fig. 9A3, wash out.. Only a long lasting depolarization persisted under CNQX ŽFig. 9A2.. A dual NMDA and non-NMDA receptor-mediated transmission therefore takes place between fibers ascending from the lateral columns and their target inhibitory interneurons. The glycine receptor antagonist strychnine Ž5 mM. was added to the Ringer’s solution to identify the neurotransmitter responsible for the inhibitory inputs. Under this condition, the inhibition elicited in reticulospinal neurons Ž n s 5. by stimulation of the lateral columns on the contralateral side was reduced and ultimately abolished ŽFig. 9B2,B3.. In all cases Žfive cells in five experiments., an
Fig. 9. Pharmacological characteristics of the inhibitory responses evoked by lateral column stimulation in reticulospinal neurons within the posterior rhombencephalic reticular nucleus. A: Synaptic responses elicited by the stimulation of the contralateral lateral columns with a train of three pulses in Mg 2q-free Ringer. Each trace is an average of 10 sweeps. A1: Control responses. Note the hyperpolarization preceding the depolarization. A2: Adding the non-NMDA receptor antagonist, CNQX Ž10 mM. to the Mg 2q-free Ringer blocked the IPSP. Only a long lasting depolarization persisted under CNQX. A3: Recovery of the responses after washing out CNQX. B: Synaptic responses elicited in another reticulospinal neuron by the stimulation of the contralateral lateral columns with a train of three pulses. B1: Control inhibitory responses. B2,B3: Adding the glycine receptor antagonist strychnine Ž5 mM. to the perfusing Ringer’s blocked the IPSPs within 10–30 min. A long lasting excitation occurred under strychnine.
290
L. Vinay et al.r Brain Research 808 (1998) 279–293
Fig. 10. Effect of the GABA B -receptor agonist, Žy.-baclofen, on synaptic responses elicited in reticulospinal neurons within the posterior rhombencephalic reticular nucleus by stimulating lateral columns. A: Synaptic responses to the stimulation of the ipsilateral lateral column ŽiLC. with a train of three pulses at 125 Hz. B: Synaptic responses to the stimulation of the contralateral side ŽcLC.. Responses were recorded in the control situation ŽA1,B1., after adding 100 mM Žy.-baclofen to the bathing Ringer’s solution ŽA2,B2. and after removing Žy.-baclofen by washing with normal Ringer ŽA3,B3..
excitation was revealed after strychnine ŽFig. 9B3., further substantiating that responses to the stimulation of lateral columns in lampreys are composed of mixed excitation and inhibition.
tory action onto the pathway from lateral columns to reticulospinal neurons.
3.5. GABAergic modulation of spinal inputs from lateral columns
This study has shown that stimulation of lateral columns on either side elicits a mixture of excitatory and inhibitory responses in reticulospinal neurons of lampreys. Excitatory amino acid transmission is responsible for the excitation. NMDA-receptor activation plays a role in the late component of the excitatory response as well as in exciting the inhibitory relay cells. Moreover, NMDA receptors are present on reticulospinal neurons. Glycine transmission is responsible for the inhibitory transmission, whilst both excitatory and inhibitory pathways are under a GABA B modulatory action.
Interestingly, Žy.-baclofen Ž100 mM., a GABA B -receptor agonist, depressed markedly the synaptic responses elicited in reticulospinal neurons by lateral column stimulations. Both excitatory and inhibitory components of the responses were abolished after Žy.-baclofen Žcompare Fig. 10A1, B1 with Fig. 10A2,B2., an effect which was reversed after washing out Žy.-baclofen ŽFig. 10A3,B3.. These results suggest the presence of a GABA B modula-
4. Discussion
L. Vinay et al.r Brain Research 808 (1998) 279–293
4.1. Ascending lateral column pathway Few studies have investigated the ascending projections from lateral columns to the brainstem of lampreys w25x. An ascending pathway referred to as the spinal lemniscus was described, which courses through the ventrolateral spinal cord and medulla. The pathway ascends to the region of the isthmus and may reach the mesencephalic tegmentum. In the caudal brainstem, ascending fibres are mainly located ventrally and laterally although some sparse fibers project medially in the reticular formation. Injections of either horseradish peroxidase in the area of the spinal lemniscus or cobalt–lysine in the PRRN reveals retrograde labeling of cells in the spinal cord w25,34x. The axons of these neurons course laterally in the spinal cord and are very likely responsible for the early responses elicited in reticulospinal neurons in the present study. These short latency responses persisted at a high frequency of stimulation ŽFig. 1D2. which suggests a mono- or at least a disynaptic linkage between the ascending system and reticulospinal neurons. Paired intracellular recordings are however now required to further test this hypothesis. In some cells, in particular PRRN neurons, the responses evoked by stimulating the ipsilateral lateral columns with a stronger current may result in a dual electrochemical transmission, which may take place between the ascending system and reticulospinal neurons. Alternatively, these components could result from an antidromic activation of descending systems which, through collaterals, could activate reticulospinal neurons. The antidromic activation of other reticulospinal neurons cannot be excluded although Dubuc et al. w15x have not been able to reveal synaptic interactions between reticulospinal neurons of homonymous and heteronymous reticular nuclei when using paired intracellular recordings of reticulospinal neurons. Vestibulospinal axons are known to make ‘en passant’ chemical and electrotonic excitatory connections onto reticulospinal cells w26x. In our present study, the monosynaptic excitation elicited in reticulospinal cells by strong lateral column stimulation might have resulted from the antidromic activation of vestibulospinal axons. Responses of reticulospinal neurons to lateral column stimulation were also characterized by a significant reduction in amplitude when the stimulating electrode was moved caudally along the cord ŽFig. 3., and a marked potentiation with a train of two-five pulses ŽFig. 1,Figs. 3–5.. A significant part of the spinal input to reticulospinal neurons, particularly from caudal levels, may therefore be conveyed through a polysynaptic pathway. Anatomical data showing that the number of spino-bulbar cells per segment indeed rapidly decreases rostro-caudally, suggested the existence of two systems ascending from the caudal spinal cord, one involving direct spino-reticular cells and another one including relay cells in the most rostral spinal segments and the caudal rhombencephalon w34x.
291
Brainstem projecting cells are located bilaterally in the spinal cord, with a contralateral dominance. The axons of contralaterally projecting cells decussate near the level of the cell body. The excitatory responses elicited in the present study by stimulation of the ipsilateral side were larger and had a shorter latency than those induced from the contralateral side. Responses from the contralateral side were most likely conveyed through a polysynaptic pathway. Transmission along this pathway required either a temporal facilitation with the use of a train of stimuli ŽFig. 1C2,C3. or a spatial facilitation of inputs with the activation of a high number of axons. This may account for the threshold intensity to evoke a response which was as a rule higher for contralateral as compared to the ipsilateral stimulation. 4.2. Transmitter systems The excitation elicited by lateral column stimulation relies primarily on excitatory amino acid transmission. After administration of CNQX and AP5, a small slow CNQX and AP5 insensitive component remained ŽFig. 5A.. Whether this is due to insufficient blockade of the excitatory amino acid transmission, or an increased transient potassium level or the involvement of another type of transmitter is not clear. The excitatory responses show an early component, which results from a non-NMDA receptor activation. Although the EPSPs recorded in normal Ringer appeared to be mediated predominantly by these receptors, the EPSPs were significantly increased in amplitude and duration in Mg 2q-free Ringer. Removing Mg 2q ions from the Ringer’s presumably unmasked the NMDA component w21,22x. This action may be on the relay interneuronal level andror on the reticulospinal cells. We demonstrate here the presence of NMDA receptors on reticulospinal neurons of the PRRN ŽFig. 7.. In a previous study, Dryer w13x reported that reticulospinal neurons in lampreys do not have NMDA receptors on their cell membrane. The experiments were, however, performed with normal concentrations of Mg 2q, which presumably masked the depolarizing effects of NMDA. A study by Alford and Dubuc w1x, has shown that disynaptic vestibulo-reticular responses also rely on NMDA receptor activation. Similar dual-component ŽNMDA and non-NMDA. glutamatergic transmission has been reported in lampreys for ascending dorsal column inputs w15x, for trigeminal inputs w29x, between descending axons and spinal neurons w3,10x as well as in other vertebrate species Žw11x; see Ref. w17x for review.. Interestingly, the excitation which is provided by the lateral columns to inhibitory interneurons also relies, at least partly, on NMDA receptor activation. The inhibition evoked in reticulospinal neurons was indeed markedly potentiated in Mg 2q-free Ringer’s ŽFig. 8.. Considering
292
L. Vinay et al.r Brain Research 808 (1998) 279–293
the voltage dependency of the NMDA receptor activation, it is likely that, as the general level of excitability increases, both excitatory and inhibitory interneurons will receive an increased synaptic drive. Such mechanism might be important in maintaining the balance of excitation and inhibition to reticulospinal neurons during locomotion. Dorsal column fibres ascending to the brainstem have also been reported to project to two populations of relays cells, an excitatory group and an inhibitory group w15,16x. While excitation from dorsal column fibres themselves and the excitatory relay cells depended on excitatory amino acid transmission, the inhibitory relay cells were glycinergic. The inhibitory component followed high frequency stimulation Ž10 Hz., although it was disynaptic. It is likely that a similar efficient glycinergic transmission is also present for lateral column inputs to reticulospinal neurons since it persists at stimulation rates of 10 Hz ŽFig. 8B2. and is abolished by strychnine ŽFig. 9B.. It therefore appears that excitatory amino acid and glycinergic transmission are common features of ascending spinal pathways to reticulospinal neurons, including the dorsal column pathways as well as the lateral funiculi. As previously reported for the dorsal column input w15x, excitatory and inhibitory inputs from lateral columns were abolished by Žy.-baclofen ŽFig. 10., suggesting that these pathways are under a modulatory action of GABA B receptors. It is not yet possible to establish the exact location of the action of GABA. Presynaptic mechanisms may be considered because GABA B receptor-mediated presynaptic inhibition has been demonstrated in the lamprey spinal cord on dorsal column afferents and interneurons w2,8x. If located at a presynaptic level, GABA B receptors may modify the gain of spinal inputs onto reticulospinal neurons, depending on the ongoing motor behavior.
ipsilateral reticulospinal neurons. If these inputs were to be phasically active during locomotion, they could be responsible for the alternation of depolarization and hyperpolarization which occurs in reticulospinal neurons during each locomotor cycle. The phasic information conveyed by the lateral column pathway is likely also responsible for the locomotor-related modulation of both the activity of vestibulospinal neurons w5x and the transmission from vestibular inputs to reticulospinal neurons w6x. Due to this modulation, supraspinal neurons may therefore respond effectively to an external stimulus during one phase of the locomotor cycle but not in another, so that the activity in descending pathways in response to an external stimulus will match the ongoing locomotor activity. Vinay and Grillner w31x showed that an intracellular stimulation of single reticulospinal neurons during the ipsilateral ventral root burst, when they were depolarized, increased the cycle duration by prolonging the ipsilateral motor burst. This mechanism may be important for equilibrium control and steering, in response to sensory or goal-directed inputs. Conversely, blocking the synaptic transmission in the brainstem, thereby eliminating the feedback from the spinal cord, was able to speed up the locomotor rhythm. More recently, Buchanan and Kasicki w4x showed that the peak to peak amplitude of membrane potential oscillations was larger in motoneurons and some interneurons during brainstem-dependent fictive swimming, as compared to Dglutamate-induced swimming in the isolated spinal cord. The enhanced oscillation amplitudes were interpreted to be due to the rhythmic descending activity from the brainstem. The lateral column pathway is the afferent link of a spino–reticulo-spinal loop w14,30–32x which can therefore be considered as an integral part of the locomotor patterngenerating circuitry.
4.3. The role of lateral column inputs during locomotion Lamprey reticulospinal neurons exhibit rhythmic membrane potential oscillations during both fictive w18,19x and active locomotion w9x. These oscillations result from alternating excitation and inhibition in each locomotor cycle with the depolarization occurring during the ipsilateral ventral root discharges and the hyperpolarization during the contralateral motor activity. Using a double-bath paradigm, Dubuc and Grillner w14x have shown that this rhythmic activity originates, at least partly, from the spinal cord locomotor networks. Vinay and Grillner w30x then recorded locomotor-related rhythmic discharges from ascending axons running in the lateral columns. Some of the axons discharged in phase with either the ipsilateral or the contralateral motor activity, others with either of the transition phases between the two bursts. In the present study, we have shown that stimulation of the lateral columns, where rhythmically active ascending axons had been recorded, elicits a mixture of excitation and inhibition in
Acknowledgements This work was supported by a Group Grant ŽNeurological Sciences. from the Canadian Medical Research Council, by the FCAR ŽQuebec Ex´ ., by the France–Quebec ´ change Program, by the Swedish Medical Research Council Žno. 3026. and by the Swedish Science Research Council Žno. 5236.. L. Vinay was supported by a grant from the ŽFrance.. The auFondation pour la Recherche Medicale ´ thors express their gratitude to H. Axegren, M. Bredmyr, G. Delforge, G. Duhau, J. Jodoin, D. Lauzier and A. Yvinec for their technical assistance and to S. Dupuis for computer programming. We would also like to express our gratitude to J.E. Gersmehl from the U.S. Fish and Wildlife Service as well as to Dr. J.G. Seelye, Mr. W.D. Swink and Mrs. M.K. Jones from the Lake Huron Biological Station for their kind supply of lampreys to the Montreal ´ laboratory.
L. Vinay et al.r Brain Research 808 (1998) 279–293
References w1x S. Alford, R. Dubuc, Glutamate metabotropic receptor mediated depression of synaptic inputs to lamprey reticulospinal neurones, Brain Res. 605 Ž1993. 175–179. w2x S. Alford, S. Grillner, The involvement of GABAb receptors and coupled G-proteins in spinal GABAergic presynaptic inhibition, J. Neurosci. 11 Ž1991. 3718–3726. w3x J.T. Buchanan, L. Brodin, N. Dale, S. Grillner, Reticulospinal neurones activate excitatory amino acid receptors, Brain Res. 408 Ž1987. 321–325. w4x J.T. Buchanan, S. Kasicki, Activities of spinal neurons during brain stem-dependent fictive swimming in lamprey, J. Neurophysiol. 73 Ž1995. 80–87. w5x N. Bussieres, R. Dubuc, Phasic modulation of vestibulospinal neu` ron activity during fictive locomotion in lampreys, Brain Res. 575 Ž1992. 174–179. w6x N. Bussieres, R. Dubuc, Phasic modulation of transmission from ` vestibular inputs to reticulospinal neurons during fictive locomotion in lampreys, Brain Res. 582 Ž1992. 147–153. w7x N. Bussieres, R. Dubuc, Reticulospinal neurons and their spinal ` projections in lampreys: a study using retrograde tracers, Soc. Neurosci. Abstr. 20 Ž1994. 1586, ŽAbstract.. w8x J. Christenson, S. Grillner, Primary afferents evoke excitatory amino acid receptor-mediated EPSPs that are modulated by presynaptic GABAb receptors in lamprey, J. Neurophysiol. 66 Ž1991. 2141– 2149. w9x S. Coghlin, N. Bussieres, ` R. Dubuc, Study of reticulospinal neurones and their sensory inputs during active and fictive swimming in a lamprey semi-intact preparation, Soc. Neurosci. Abstr. 18 Ž1992. 314, ŽAbstract.. w10x N. Dale, S. Grillner, Dual-component synaptic potentials in the lamprey mediated by excitatory amino acid receptors, J. Neurosci. 6 Ž1986. 2653–2661. w11x N. Dale, A. Roberts, Dual-component amino acid-mediated synaptic potentials: excitatory drive for swimming in Xenopus embryos, J. Physiol. 363 Ž1985. 35–59. w12x G.R.J. Davis, A.D. McClellan, Extent and time course of restoration of descending brainstem projections in spinal cord-transected lamprey, J. Comp. Neurol. 344 Ž1994. 65–82. w13x S.E. Dryer, Excitatory amino acid-evoked membrane currents and excitatory synaptic transmission in lamprey reticulospinal neurons, Brain Res. 443 Ž1988. 173–182. w14x R. Dubuc, S. Grillner, The role of spinal cord inputs in modulating the activity of reticulospinal neurons during fictive locomotion in the lamprey, Brain Res. 483 Ž1989. 196–200. w15x R. Dubuc, F. Bongianni, Y. Ohta, S. Grillner, Dorsal root and dorsal column mediated synaptic inputs to reticulospinal neurons in lampreys: involvement of glutamatergic, glycinergic and GABAergic transmission, J. Comp. Neurol. 327 Ž1993. 251–259. w16x R. Dubuc, F. Bongianni, Y. Ohta, S. Grillner, Anatomical and physiological study of brainstem nuclei relaying dorsal column inputs in lampreys, J. Comp. Neurol. 327 Ž1993. 260–270. w17x P.M. Headley, S. Grillner, Excitatory amino acids and synaptic transmission: the evidence for a physiological function, T.I.P.S. 11 Ž1990. 205–211.
293
w18x S. Kasicki, S. Grillner, Muller cells and other reticulospinal neu¨ rones are physically active during fictive locomotion in the isolated nervous system of the lamprey, Neurosci. Lett. 69 Ž1986. 239–243. w19x S. Kasicki, S. Grillner, Y. Ohta, R. Dubuc, L. Brodin, Phasic modulation of reticulospinal neurones during fictive locomotion and other types of spinal motor activity in lamprey, Brain Res. 484 Ž1989. 203–216. w20x R.J. Martin, A study of the morphology of the large reticulospinal neurons of the lamprey amnocoete by intracellular injection of procion yellow, Brain Behav. Evol. 16 Ž1979. 1–18. w21x M.L. Mayer, G.L. Westbrook, P.B. Guthrie, Voltage-dependent block by Mg 2q of NMDA responses in spinal cord neurones, Nature 309 Ž1984. 261–263. w22x L. Nowak, P. Bregestovski, P. Ascher, A. Herbet, A. Prochiantz, Magnesium gates glutamate-activated channels in mouse central neurones, Nature 307 Ž1984. 462–465. w23x G.N. Orlovsky, Cerebellum and locomotion, in: M. Shimamura, S. Grillner, V.R. Edgerton ŽEds.., Neurobiological Basis of Human Locomotion, Japan Scientific Societies Press, Tokyo, 1991, pp. 187–199. w24x C. Perret, Neural control of locomotion in the decorticate cat, in: R.M. Herman, S. Grillner, P.S.G. Stein, D.G. Stuart ŽEds.., Neural Control of Locomotion, Plenum, New York, 1976, pp. 587–615. w25x M. Ronan, R.G. Northcutt, Projections ascending from the spinal cord to the brain in petromyzontid and myxinoid agnathans, J. Comp. Neurol. 291 Ž1990. 491–508. w26x C.M. Rovainen, Electrophysiology of vestibulospinal and vestibuloreticulospinal systems in lampreys, J. Neurophysiol. 42 Ž1979. 745–766. w27x G.P. Swain, J.A. Snedeker, J. Ayers, M.E. Selzer, Cytoarchitecture of spinal-projecting neurons in the brain of the larval sea lamprey, J. Comp. Neurol. 336 Ž1993. 194–210. w28x G.P. Swain, J. Ayers, M.E. Selzer, Metamorphosis of spinal-projecting neurons in the brain of the sea lamprey during transformation of the larva to adult: normal anatomy and response to axotomy, J. Comp. Neurol. 362 Ž1995. 453–467. w29x G. Viana Di Prisco, Y. Ohta, F. Bongianni, S. Grillner, R. Dubuc, Trigeminal inputs to reticulospinal neurones in lampreys are mediated by excitatory and inhibitory amino acids, Brain Res. 695 Ž1995. 76–80. w30x L. Vinay, S. Grillner, Spino-bulbar neurons convey information to the brainstem about different phases of the locomotor cycle in the lamprey, Brain Res. 582 Ž1992. 134–138. w31x L. Vinay, S. Grillner, The spino–reticulo-spinal loop can slow down the NMDA-activated spinal locomotor network in lamprey, NeuroReport 4 Ž1993. 609–612. w32x L. Vinay, S. Grillner, The spino–reticulo-spinal loop and the control of locomotion in lamprey, Eur. J. Neurosci. Suppl. 6 Ž1993. 19, ŽAbstract.. w33x L. Vinay, Y. Padel, D. Bourbonnais, H. Steffens, An ascending spinal pathway transmitting a central rhythmic pattern to the magnocellular red nucleus in the cat, Exp. Brain Res. 97 Ž1993. 61–70. w34x L. Vinay, N. Bussieres, O. Shupliakov, R. Dubuc, S. Grillner, ` Anatomical study of spino-bulbar neurons in lampreys, J. Comp. Neurol. 397 Ž1998. 475–492. w35x W.O. Wickelgren, Physiological and anatomical characteristics of reticulospinal neurones in lamprey, J. Physiol. 270 Ž1977. 89–114.